Cloud-fraud-monitoring

A backend system built with Java Spring Boot and AWS to detect and flag fraudulent transactions in real-time using rule-based logic. It integrates Kafka for scalable message streaming, MySQL (AWS RDS) for persistent storage, and Power BI for fraud analytics and reporting.

Tech Stack

Java 17 Spring Boot Kafka MySQL (AWS RDS) Docker AWS (CloudWatch, RDS, EC2) Power BI

Features

✅ REST API to handle transaction data (GET, POST, DELETE) ✅ Real-time fraud detection using rule-based logic ✅ Kafka integration for transaction event streaming ✅ Power BI dashboard connected to live AWS RDS database ✅ Dockerized microservice architecture ✅ Postman collection for API testing

Fraud Detection Rules

Transaction amount > $1000 Multiple transactions from same user in a short time (60 sec) New location or device First transaction is unusually large Suspicious device pattern (e.g., "Chrome on Ubuntu Server") [Optional future rule] Integration with ML-based detection

Dashboard (Power BI)

Pie chart: Fraudulent vs Non-Fraudulent transactions Filters by date, user ID, location Auto-refreshes from AWS RDS Future scope: Device-based drilldowns

How to Run the Project (Dev Mode)

Step 1: Build the Spring Boot Application

./mvnw clean install

Step 2: Run the App and Kafka using Docker Compose

docker compose down # Stop and remove previous containers (optional clean start) docker compose up --build # Build and run all services (Kafka, MySQL, Spring Boot)